Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!uakari.primate.wisc.edu!gem.mps.ohio-state.edu!apple!bbn!news
From: news@bbn.COM (News system owner ID)
Newsgroups: comp.dcom.modems
Subject: Re: Terminal Programs for 9600 baud, v.32 MNP 5
Message-ID: <46961@bbn.COM>
Date: 16 Oct 89 15:01:11 GMT
References: <20551@usc.edu>
Reply-To: pplacewa@antares.bbn.com (Paul W. Placeway)
Organization: Bolt Beranek and Newman Inc., Cambridge MA
Lines: 75

(I'm posting a copy of this because it may be of use to other Mac
users out there -- PWP)

burke@pollux.usc.edu (Sean Burke) writes:
<   Let's talk about modem programs.  My modem program is Red Ryder 10.3, and
< I'm unhappy with its performance on a number of points.

Well, that was your first mistake...

<   - RR cannot keep up with the data in terminal mode.  The screen doesn't 
<     scroll very much faster than at 2400 baud.  If I turn off Xon/Xoff at the
<     receiving end, much stuff gets lost.  By contrast,  MacKermit's scrolling
<     seems to be full speed.

Yup.  I suspect that the problem with RR is that it tries to process
each character one at a time.  There is a fairly sucessful hack that
Mac Kermit uses to pick up some speed: it buffers normal characters
(when not doing an escape sequence) until it either sees a control
character (that is, non-normal), has to autowrap, or runs out of
chacters to display.  Then it writes them all out as a string.  This
method works quite well, as long as all the conditions to flush are
right.

<   - File transfers are incredibly slow.  Xmodem ran at about 140cps, 
<     regardless of whether I use the Xmodem spoofing or not.  By contrast,
<     if I use Freeterm with Xmodem spoofing, I get the highest transfer rate
<     yet, a screaming 760cps, but spoofing only works when receiveing.  Kermit 
<     did better, perhaps because I could set 1K packets.  RR and MacKermit both
<     got into the 250cps range.

Try running kermit (protocol) with a large packet size: set the
recieve packet size on both sides to about 950 bytes (1000 works
_slightly_ better, but 950 is a safer guess).  This should improve
things somewhat.

The other problem you are probably running into is a delay between
when Kermit is done sending a packet, and when MNP decides to actually
send it across the wire.  The best work-around I have to that is to
(gulp) turn off MNP.  Obviously there should be a better way -- I
don't know if there is any way to

<   - What I would really like to do is to use speed conversion option, with
<     the serial port set at 19.2kbd, so that I can take advantage of the
<     compression provided by MNP-5.  None of the terminal programs I can
<     scrape up offer this option, though the modem manual mentions a program
<     called MAcKNOWLEDGE.

Mac Kermit 0.98(62) does this.  FTP a copy from
watsun.cc.columbia.edu; it's /kermit/test/ckmker.hqx.  Don't use
0.97(57) -- it had some font-switching bugs.

<   So you say, why not use Freeterm?  Well, Freeterm doesn't do v100
< emulation.  MacKermit does, and it's starting to look like my
< terminal program of choice, if only because the screen will update
< at full speed. 

Retorical question: why use Freeterm?  What does it do that Mac Kermit
doesn't (I've never used it)?

<   As long as I'm bitching about RR, I've just got to mention that I _STILL_
< haven't been able to turn off that stupid title screen, and I PAID for this
< program!  And I've never been able to get Xmodem or Kermit to use 1K packets.

Well, I think this is a reverse version of you get what you pay for.
RR costs and is low quality; Mac Kermit, Freeterm (from what I've
heard), Zterm, and a few others are free and high quality.

<   So, in summary, V.32 is nice, and you can get it for a reasonable price 
< these days, but you're going to have to work to take advantage of the new
< technology.

I'm glad to hear this.  Maybe I'll get a v.32 soon too....  :-)

		-- Paul Placeway <PPlaceway@bbn.com>
		   Mac Kermit coord. (among other things)